"Second Source" Confirms Project Veritas Report on YouTube Meddling in Irish Abortion

OMM 0000 · 28-06-2019 07:02AM

Walter Bishop wrote: »

They have been 100% civil, you just don't like being challenged it seems. If you get this defensive the first time someone questions your alleged authority...

They're making up things (pretending I said x when I said y), saying I don't know what I'm talking about, saying I'm motivated to treat people like idiots, saying I'm lying about my job and background, and generally being very aggressive and trying to argue about something which doesn't appear to be their area of work.

Normal people don't talk or act like this. It's bizarre.

acorntoast · 28-06-2019 07:06AM

Deleted User wrote: »

The main goal of the algorithms of service providers like You Tube is not simply to police the content of the articles as your simplistic narrative attempted to imply.

OMM 0000 wrote:

I'm surprised to hear I was saying the machine learning used by YouTube is "simply to police the content of the articles".

Deleted User wrote: »

No one said you said that. You are making this up now.

Busted.

28-06-2019 08:03AM

acorntoast wrote: »

Busted.

Busted yourself. Very selective quoting there from you. Allow me to improve it for you:

Deleted User wrote: »

The main goal of the algorithms of service providers like You Tube is not simply to police the content of the articles as your simplistic narrative attempted to imply.

Deleted User wrote: »

No one said you said that. You are making this up now.

What I very clearly said - and anyone can scroll up to see this - was that I said the simplistic way you described the process builds that _implication_.

Do try harder.

rossie1977 · 28-06-2019 10:18AM

I guess Donegal doesn't get youtube..

The US far right media should have done a little research on what actually influenced the majority of people, it certainly wasn't youtube, it was death of Savita Halappanavar

OMM 0000 · 28-06-2019 01:19PM

Google talking a little bit about how their staff's bias affects their machine learning models:

https://www.standard.co.uk/tech/bias-in-ai-google-head-of-ethical-machine-learning-a3986256.html

seamus · 28-06-2019 01:38PM

OMM 0000 wrote: »

Google talking a little bit about how their staff's bias affects their machine learning models:

https://www.standard.co.uk/tech/bias-in-ai-google-head-of-ethical-machine-learning-a3986256.html

That's misrepresentation.

You're implying that the staff's personal biases are injected into the machine learning models, when that's not the case, it's right there in the article.

Machine learning and AI is entirely dependent on the nature and quality of its data input.

So if the software is developed and trained in San Francisco by predominantly white male developers, then much of what the machine "knows" about the world will be culturally skewed towards the California middle class.
Which is in fact exactly what you would expect, because that's how humans learn. If you grow up a white Irish middle class male, then you will be culturally biased in that way. But we didn't expect this to happen - maybe we thought we'd be worse at AI than we actually are - and so now we have to tweak the inputs to eliminate local bias from the machine.

Your implication is that staff themselves are deliberately "telling" the machines what to think (or not to think). When that's not the case.

On the subject of your OP, it may very well be that pro-choice material made its way to the top of search rankings. But that's the because the mchine learning algorithms determined that people in Ireland predominantly wished to see pro-choice material more than anti-abortion, and so that's how it weighted the search results.

Which is not a case of google search influencing the outcome of referendums, in fact it's the exact opposite - the mood on the referendum influenced the rankings in google search.

OMM 0000 · 28-06-2019 02:01PM

seamus wrote: »

That's misrepresentation.

You're implying that the staff's personal biases are injected into the machine learning models, when that's not the case, it's right there in the article.

Machine learning and AI is entirely dependent on the nature and quality of its data input.

So if the software is developed and trained in San Francisco by predominantly white male developers, then much of what the machine "knows" about the world will be culturally skewed towards the California middle class.
Which is in fact exactly what you would expect, because that's how humans learn. If you grow up a white Irish middle class male, then you will be culturally biased in that way. But we didn't expect this to happen - maybe we thought we'd be worse at AI than we actually are - and so now we have to tweak the inputs to eliminate local bias from the machine.

Your implication is that staff themselves are deliberately "telling" the machines what to think (or not to think). When that's not the case.

On the subject of your OP, it may very well be that pro-choice material made its way to the top of search rankings. But that's the because the mchine learning algorithms determined that people in Ireland predominantly wished to see pro-choice material more than anti-abortion, and so that's how it weighted the search results.

Which is not a case of google search influencing the outcome of referendums, in fact it's the exact opposite - the mood on the referendum influenced the rankings in google search.

We're almost on the same page. In fact, you disagree with me then agree with me, so I'm a little confused by your reply.

A training set can be biased in many ways. (The article talks about this).

A model created by a left wing person is going to look different to a model created by a right wing person. This is due to their bias - the data they choose, how they test the model, what they consider an acceptable output, and so on.

I think this is natural and is very hard to solve. We (my workplace) have this problem with our own models.

Where this all gets a bit shady and possibly a bit conspiracy theory sounding is if they are purposefully adding bias. If you look at the recent undercover video of the Google Executive, she says it's up to companies like Google who have the power to prevent something like Trump happening again. That sounds like manipulating their search results, etc.

notobtuse · 28-06-2019 02:35PM

OMM 0000 wrote: »

We're almost on the same page. In fact, you disagree with me then agree with me, so I'm a little confused by your reply.

A training set can be biased in many ways. (The article talks about this).

A model created by a left wing person is going to look different to a model created by a right wing person. This is due to their bias - the data they choose, how they test the model, what they consider an acceptable output, and so on.

I think this is natural and is very hard to solve. We (my workplace) have this problem with our own models.

Where this all gets a bit shady and possibly a bit conspiracy theory sounding is if they are purposefully adding bias. If you look at the recent undercover video of the Google Executive, she says it's up to companies like Google who have the power to prevent something like Trump happening again. That sounds like manipulating their search results, etc.

Reports indicate Google employees donate to Democrats over Republicans by a almost 10 to 1 margin. If what you state about employee bias is true then Google is highly biased against Republicans and it corresponds to Republican or right leaning sources being highly targeted negatively.

Annasopra · 28-06-2019 03:06PM

OMM 0000 wrote: »

I think it's important to read all sorts of news sources. There is so much misinformation by everyone, you have to read it all and come to your own conclusions.

Do you find it hard to believe Google could try to influence search results?

Nah. Breitbart can't be trusted at all. It's a far right fascist racist publication.

notobtuse · 28-06-2019 03:25PM

Joeytheparrot wrote: »

Nah. Breitbart can't be trusted at all. It's a far right fascist racist publication.

Allsides.com is a highly trusted site. They present news articles on a topic from sources on the Left, Center and Right... a very balanced approach. The list has Breitbart on the far right, but it is still included because it can be a legitimate source for news. If I used your logic we can say The New Yorker, MSNBC, The Huffington Post, Slate, Mother Jones, and The Nation can’t be trusted at all because of their radical ultra-leftism.

AllSidesMediaBiasChart-Version1.1_0.jpg

rossie1977 · 28-06-2019 05:17PM

NY times opinion pieces can be right or left so that chart isn't correct
https://www.nytimes.com/2018/07/06/opinion/democratic-socialism-alexandria-ocasio-cortez.html
https://www.nytimes.com/2017/07/06/opinion/center-democrats-identity-politics.html

OMM 0000 · 28-06-2019 05:23PM

rossie1977 wrote: »

NY times opinion pieces can be right or left so that chart isn't correct
https://www.nytimes.com/2018/07/06/opinion/democratic-socialism-alexandria-ocasio-cortez.html
https://www.nytimes.com/2017/07/06/opinion/center-democrats-identity-politics.html

Just because the NY Times occasionally has some opinion pieces which aren't left wing doesn't make it centrist or right wing. It's absolutely a left leaning publication.

Fox News has guests who are left wing, sometimes bat **** crazy far left, but no one for a second thinks Fox News is centrist or left wing.

suicide_circus · 30-06-2019 07:52AM

i dont read Breitbart but i would just see them as a counterpoint to the guardian or any of the establishment/liberal media sources which all beat the same drum. no harm with a bit of diversity of opinion.

B0jangles · 30-06-2019 08:07AM

suicide_circus wrote: »

i dont read Breitbart but i would just see them as a counterpoint to the guardian or any of the establishment/liberal media sources which all beat the same drum. no harm with a bit of diversity of opinion.

Diversity of opinion is good, but if a story is shown to be based on false information, the Guardian will print a retraction, Breitbart will just keep pushing it regardless. It's not about political stance, it's about whether or not the source thinks it is important that news reporting be based on facts.

suicide_circus · 30-06-2019 08:12AM

B0jangles wrote: »

Diversity of opinion is good, but if a story is shown to be based on false information, the Guardian will print a retraction, Breitbart will just keep pushing it regardless. It's not about political stance, it's about whether or not the source thinks it is important that news reporting be based on facts.

fair enough but i wouldn't trust either outlet not to give me a heavily biased account of a story.

30-06-2019 08:16AM

B0jangles wrote: »

Diversity of opinion is good, but if a story is shown to be based on false information, the Guardian will print a retraction, Breitbart will just keep pushing it regardless. It's not about political stance, it's about whether or not the source thinks it is important that news reporting be based on facts.

Just googled the Guardian along with Trump's quote.

June 2015: "They're bringing drugs. They're bringing crime. Their rapists. And some I assume are good people."

July 2015: "They're bringing drugs. They're bringing crime. They're rapists. And some I assume are good people."

Even the Guardian isn't above blatant fake news along with the rest of the media. They started with the correct transcription and moved to the incorrect inflammatory one. I don't think they've retracted and apologised.

Google the second one and see what news outlets come up. It's all of them.

B0jangles · 30-06-2019 08:22AM

suicide_circus wrote: »

fair enough but i wouldn't trust either outlet not to give me a heavily biased account of a story.

But editorial stance aside, in papers like the Guardian you can be fairly confident that the core facts of a story will be accurate and if they aren't, they will quickly print a retraction and a correction.

Breitbart won't, they will happily, intentionally print things which are entirely false and won't admit it, except maybe if they are about to be sued by someone very rich.
If you look up retractions and The Guardian you get this:

https://www.theguardian.com/theguardian/series/correctionsandclarifications
a continuous public display of every correction no matter how minor

if you look up retractions and Breitbart the most recent seems to be from 2016 - in three years they've only corrected themselves once, and before that the next most recent is I think, 2011.

That is not a good sign.

even to some fellow conservatives, Breitbart is the epitome of fake news
https://www.theamericanconservative.com/dreher/breitbart-fake-news-alex-marlow/comment-page-2/

B0jangles · 30-06-2019 08:25AM

Deleted User wrote: »

Just googled the Guardian along with Trump's quote.

June 2015: "They're bringing drugs. They're bringing crime. Their rapists. And some I assume are good people."

July 2015: "They're bringing drugs. They're bringing crime. They're rapists. And some I assume are good people."

Even the Guardian isn't above blatant fake news along with the rest of the media. They started with the correct transcription and moved to the incorrect inflammatory one. I don't think they've retracted and apologised.

Google the second one and see what news outlets come up. It's all of them.

How could anyone tell from a spoken word whether or not someone was saying 'their' or 'they're'? I'm not even sure which one you think it the good one, and which is the bad one, they are both pretty terrible.

30-06-2019 08:27AM

B0jangles wrote: »

How could anyone tell from a spoken word whether or not someone was saying 'their' or 'they're'? I'm not even sure which one you think it the good one, and which is the bad one, they are both pretty terrible.

Because it's extremely obvious when you watch the video. And because "they're" makes no sense whatsoever considering what he had just said, and what came after.

"They're" is way worse and is a racist thing to say. "Their" isn't.

B0jangles · 30-06-2019 08:30AM

Deleted User wrote: »

Because it's extremely obvious when you watch the video. And because "they're" makes no sense whatsoever considering what he had just said, and what came after.

Are you telling me you can distinguish between the homophones 'They're' and 'their'? Like, they sound exactly the same? And the sentence makes perfect sense, if the intention is to claim that the majority of mexican immigrants are dangerous criminals.

30-06-2019 08:35AM

B0jangles wrote: »

Are you telling me you can distinguish between the homophones 'They're' and 'their'? Like, they sound exactly the same? And the sentence makes perfect sense, if the intention is to claim that the majority of mexican immigrants are dangerous criminals.

Of course I can distinguish the difference. Listen to the below at 20 seconds in and tell me which he said.

It isn't a standalone sentence. It's the third example of things he says they're bringing.

B0jangles · 30-06-2019 08:41AM

Deleted User wrote: »

Of course I can distinguish the difference. Listen to the below at 20 seconds in and tell me which he said.

It isn't a standalone sentence. It's the third example of things he says they're bringing.

He said either 'They're' or 'Their' or 'There'. Because they are homophones. That means that even though the written word is different, the spoken word is identical.

It surely makes more sense in a rhetorical sense for it to be the same 'They're' that he just used twice in a row. Repetition is a key element in effective speaking. He's hammering home his central point that mexican immigrants are a critical threat to the US. It's one he makes on a very regular basis, I do not know why this particular instance is such a reach for you.

30-06-2019 08:52AM

B0jangles wrote: »

He said either 'They're' or 'Their' or 'There'. Because they are homophones. That means that even though the written word is different, the spoken word is identical.

It surely makes more sense in a rhetorical sense for it to be the same 'They're' that he just used twice in a row. Repetition is a key element in effective speaking. He's hammering home his central point that mexican immigrants are a critical threat to the US. It's one he makes on a very regular basis, I do not know why this particular instance is such a reach for you.

He said "their". I've been an English teacher for nearly a decade and not a single student would ever transcribe that as "they're". No one would. My girlfriend is a translator / transcriber and she agreed it's definitely "their".

Anyways, we can move on. Go back to giving your honest and unbiased critique of the media.

B0jangles · 30-06-2019 08:59AM

Deleted User wrote: »

He said "their". I've been an English teacher for nearly a decade and not a single student would ever transcribe that as "they're". No one would. My girlfriend is a translator / transcriber and she agreed it's definitely "their".

Anyways, we can move on. Go back to giving your honest and unbiased critique of the media.

An appeal to your own authority as a teacher, plus the reported support of your students and your girlfriend does not alter the fact that the two words in question are homophones. An english teacher really should understand what a homophone is.

30-06-2019 09:06AM

B0jangles wrote: »

An appeal to your own authority as a teacher, plus the reported support of your students and your girlfriend does not alter the fact that the two words in question are homophones. An english teacher really should understand what a homophone is.

Of course I know what they are.

But it's irrelevant. In the Trump video, he didn't say "they're" because it wouldn't have sounded like that. And it was the third example in a row. And he wouldn't say "They're rapists but some are good people."

If your argument is that people can never tell the difference between homophones, then I don't really know what else to say.

B0jangles · 30-06-2019 09:14AM

Deleted User wrote: »

Of course I know what they are.

But it's irrelevant. In the Trump video, he didn't say "they're" because it wouldn't have sounded like that. And it was the third example in a row. And he wouldn't say "They're rapists but some are good people."

Look, you've clearly made your mind up about this, which is absolutely your right. But there is no evidence to support your opinion. The two words are identical when spoken. Context is the only way you can tell them apart. You prefer one interpretation, an awful lot of other people think the other interpretation makes perfect sense given Trump's history of inflammatory and unsupported claims about the imminent threat immigrants pose.

But this is a thread about project veritas and the irish abortion referendum, which has widened into bias in the media. If you want to keep talking about a clip from a Trump rally from over 4 years ago, you should probably bring it to a more appropriate thread.

30-06-2019 09:20AM

I said we could move on a few posts back. If you think it was you yourself I wanted to convince, you're mistaken. It was your credibility I was after.

So now, we can move on.

B0jangles · 30-06-2019 09:31AM

That seems very weird and petty, but ok?

batgoat · 30-06-2019 11:08AM

According to that article even “repeal the 8th” was blacklisted which would be as likely to hit pro choice results as well as anti choice.... Inclined to say Breitbart invented a vague story and that's about it...

Faugheen · 30-06-2019 11:52AM

People are actually asking for evidence that Breitbart is biased?

Jesus wept.

"Second Source" Confirms Project Veritas Report on YouTube Meddling in Irish Abortion

Comments